摘要
提出了一种基于差分图像运动检测和轮廓提取的跟踪与分割图像的方法。该方法首先从初始图像中提取目标的轮廓,然后利用相邻帧之间的差分图像初步确定目标在每帧图像中的粗略位置,最后把从上一帧图像中得到的目标轮廓置于该位置,并作为轮廓提取的初始值,由此可得到对目标的准确分割。
A method for tracking of moving objects was presented. The developed algorithm was based on contour extraction and change detection in image sequences. This approach extracted the contour of the moving object from the first image, and located objects approximately through the change between the neighboring frames. In each frame, the precise segmetation is implemented through contour extraction, while the contour extracted from the last frame is put at the approximate position as the initialization. The algorithm is tested on several real image sequence.
出处
《光学技术》
EI
CAS
CSCD
北大核心
2005年第4期565-567,共3页
Optical Technique
关键词
目标跟踪与分割
差分图像
轮廓提取
motion tracking and segmentation
change detection
contour extraction
